Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
Not open for further replies.

MacVidCards

Suspended
Nov 17, 2008
6,096
1,056
Hollywood, CA
it's the one actual shill on this entire forum that goes around accusing everybody else of having a vested interest in what they type.

that's known as projecting
https://en.wikipedia.org/wiki/Psychological_projection

I'm going to guess that you don't understand what a"shill" is.

Take a moment and look it up.

As one of the very few on here who posts as myself, I can't be one

A shill is someone pretending to be an honest Joe with GTX980 who posts 5-10 "Isn't AMD the greatest thing since sliced bread?" Posts with links to other anonymous nobodies posting similar stuff.

Don't you have some GPU upgrades to find or screws to sort?
 
  • Like
Reactions: tuxon86

goMac

macrumors 604
Apr 15, 2004
7,663
1,694
Nvidia didn't want to start competing with Fury, so they set instead GTX1080 price higher, and both can enjoy the quietness in their price categories.

So I guess we're back to hoping for Fury Nano in the Mac Pro? If Fury Nano gets under clocked too low any advantage over Polaris will evaporate though. :confused:
 

AidenShaw

macrumors P6
Feb 8, 2003
18,667
4,677
The Peninsula
I think I already said too much, on this topic...
OMG, we're really tired of this faux "oops, I said too much" nonsense. You pretend to have inside sources, but you don't. (If you did - you'd be cut off for violating NDA.)

Jon Snow - you know nothing. You were blind-sided by GP100, and now again by GP104.

I can't wait to see you back-track when Apple announces GP104-equipped Mac Pros in four weeks.
 

Dean Yu

macrumors regular
Mar 12, 2016
162
134
Toronto, Canada
Anyone knows Fury Nano's size (silicon)?
Here.
Don't worry about PCB since Apple will still use their beautiful PowerPC G3 750/ G4 7445/ G5 970 daughterboard compatible Multipin connector.... :)
 

Attachments

  • image.jpg
    image.jpg
    820.1 KB · Views: 121

goMac

macrumors 604
Apr 15, 2004
7,663
1,694
If an NDA agreement prevents you from disclosing the source of the information, it would also prevent you from disclosing the information itself.

Heh. I don't know if he specifically is an insider, but I do know there are insiders on this forum who post things, so I wouldn't rule it out. Information doesn't always come with NDAs either.
 

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853
OMG, we're really tired of this faux "oops, I said too much" nonsense. You pretend to have inside sources, but you don't. (If you did - you'd be cut off for violating NDA.)

Jon Snow - you know nothing. You were blind-sided by GP100, and now again by GP104.

I can't wait to see you back-track when Apple announces GP104-equipped Mac Pros in four weeks.
As I already said before, everything about Nvidia I post comes from my analysis of the architecture and information placed all over the place on the internet.

Things like design of future computers, and computing overall, AMD's approach for making CUDA-like API from Mantle and things that are connected to this - those are information that come from very good source. He/she does not give me specifics, only general idea, that can be extrapolated later.

And no Aiden. There will not be GP104 in Mac Pro. Apple will use AMD GPUs. I do not know which ones. Forget about Nvidia on Apple platform. It is done deal, for upcoming several years. Blame Nvidia and their try to lawsuit Apple for intellectual property that Nvidia failed to force even on Samsung patents.

P.S. I was not wrong with GP104. You just resist, the fact, that only thing that was wrong in my analysis was core clocks. TDP, core counts for both GPUs were pretty much in line with what we have seen on the presentation. But you know. Si tu parle, I could be wrong.
 

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853
Mago, you want to know the whole interposer or only die size? Die size is 596 mm2. Interposer is around 950mm2 IF I remember correctly.

P.S. GTX 1080 is 10-15% faster in DX12 games than... Fury X ;). Pretty damn good.
 

Mago

macrumors 68030
Aug 16, 2011
2,789
912
Beyond the Thunderdome
Mago, you want to know the whole interposer or only die size? Die size is 596 mm2. Interposer is around 950mm2 IF I remember correctly.

P.S. GTX 1080 is 10-15% faster in DX12 games than... Fury X ;). Pretty damn good.
Polaris xt its suposedly 232mm near 1/3 so it could have the same transistor and few more than fury nano even higher clock and still match 100w TDP
 

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853
Polaris xt its suposedly 232mm near 1/3 so it could have the same transistor and few more than fury nano even higher clock and still match 100w TDP
With HBM - yes, maybe. With GDDR5 and GDDR5X - no chance.
 

Zarniwoop

macrumors 65816
Aug 12, 2009
1,038
760
West coast, Finland
Things at GPU market are now pretty interesting. And it would turn out to be very interesting, if AMD releases along Polaris 11 and 10 also refined Fiji, Radeon Fury 2016 edition with TDP ~190W and DP1.3 and HDMI 2.0 - and still keep the price at $499 level. Similar refine as was 290 --> 390. Sur it would be still 4GB HBM and no match to GTX 1080, but it would make people think twice and have at least some sort of competition.

But if it's just Polaris 11 and 10, Nvidia has pretty much monopoly on GTX 1080 segment. And that is not bad. AMD will most likely take mobile market for six month at least with Polaris 11.
 

Zarniwoop

macrumors 65816
Aug 12, 2009
1,038
760
West coast, Finland
What comes to Radeon Fury/Fiji in Mac Pro, I still don't think it will happen. DP 1.2 and HDMI 1.4 native (HDMI 2.0 is possible with adapter). If these are not updated, no change that Apple is interested to use it in any product. Unless maybe as a TB3 external box a la Radeon Pro.

Polaris10 / GTX 1070 level will be just what their current line will use. But of course, if there's coming totally redesigned Mac's.. with different thermal needs..

Intel iGPU doesn't support Adative sync, so there might be more demand for Polaris 11 than just a power. DP 1.3/HDMI 2.0b/Adaptive sync and DSP. I know that Siri would benefit a lot of DSP, if there was any available. Currently, not.
 
Last edited:

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853
Then there could be no true "oops - I said too much."
LOL. You quoted post from first page when we were talking about AMD GCN GPUs for next Mac Pro. Where did you get information about Nvidia there? It was info, that I extrapolated from what has been told me: "multiply the performance of current GPUs on smaller package, and Thermal Envelope, from previous generation".

And it was pretty darn close to the reality. Again. My source does not know the exact config of next MP, it just knows what is happening in few areas, here and there. And yes, it had these GPUs in hands.

Where I was talking in that posts about Nvidia?
 

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853
Two possibilities, either their hardware is so strong and performant that it will defend itself, or... their Marketing department completely sucks.

One thing is waiting to adjust clocks on their hardware to counter what your competitor has to offer, second thing is being completely silent, and letting all the attention go to your competitor...
 

Mago

macrumors 68030
Aug 16, 2011
2,789
912
Beyond the Thunderdome
I just realize Apple's strategy on compute and hpc, they Trojan Horse its Swift 3.0

Those familiar with python and c# and learning Swift should have noted very similar constructs and syntax differentiated mostly by trivial things.

Ok past week was finalized Swift 3.0 specification, guess that is not compatible with previous versions and most code will need a huge make up to migrate to swift 3, but still possible some automated tool will do, is that a bad thing you may think yes that is bad, unless you're familiar with python structures, the shinny new Swift 3.0 will allow migrate python code with mere trivial changes most can be done by common refactoring but certainly it's possible to write a python -> swift 3 migration tool fully automatic and 100% precise what you get on python its what you'll get on swift 3, you may think ok we now have tons of python code to migrate to OSX iOS apps , what it has to do with HPC? , you miss Swift since 2.0 its opensource? Yes you can run swift 3 code on a cheap raspberry pi zero as well as on the biggest cluster, with few solid advantages over python given swift its fully compiled its efficiency its superior to what you can get from python compilers as pypi and close or better than raw C, plus tons of new features plus an powerful ide (Xcode also Kdevelop), imagine you are on R&D at some hexascale compute project , you can code and test it on a Mac pro importing/translating whatever code you need from C/C++/Objective-C/Python and maybe later C# too, and later with simple script migrate it to production clusters running Linux or BSD or whatever where you have the open-source Swift 3 compiler.

[Insert Tim Cook voice in keynote mode]
That's Awwwnmaaaaaahhhzzinnnnng
[End of Tim Cook voice keynote mode]
 
Last edited:

ManuelGomes

macrumors 68000
Original poster
Dec 4, 2014
1,617
354
Aveiro, Portugal
Synchro3, I was referring to the NVidia cards that are apparently not DP1.4 certified yet. I find it strange that if NVidia hasn't been able to, how would those you mentioned are? This could also be marketing work, they are not now but maybe when in fact they are released. The cards aren't even put to buy so how can you be sure?
1080 in general will be DP1.4, when it's fully released maybe, but as far as I know it is not yet the case.
We'll have to wait and see when they come.
[doublepost=1462803639][/doublepost]I was gonna say that AMD might wait a few days and reveal some more info but it seems confirmaed that Macau will be the presentation place later this month.
With availability in june this makes sense. Let's see if there will be cards out soon after.
With scarce availability of GDDR5X I have my doubts for both camps.
 

tuxon86

macrumors 65816
May 22, 2012
1,321
477
I just realize Apple's strategy on compute and hpc, they Trojan Horse its Swift 3.0

Those familiar with python and c# and learning Swift should have noted very similar constructs and syntax differentiated mostly by trivial things.

Ok past week was finalized Swift 3.0 specification, guess that is not compatible with previous versions and most code will need a huge make up to migrate to swift 3, but still possible some automated tool will do, is that a bad thing you may think yes that is bad, unless you're familiar with python structures, the shinny new Swift 3.0 will allow migrate python code with mere trivial changes most can be done by common refactoring but certainly it's possible to write a python -> swift 3 migration tool fully automatic and 100% precise what you get on python its what you'll get on swift 3, you may think ok we now have tons of python code to migrate to OSX iOS apps , what it has to do with HPC? , you miss Swift since 2.0 its opensource? Yes you can run swift 3 code on a cheap raspberry pi zero as well as on the biggest cluster, with few solid advantages over python given swift its fully compiled its efficiency its superior to what you can get from python compilers as pypi and close or better than raw C, plus tons of new features plus an powerful ide (Xcode also Kdevelop), imagine you are on R&D at some hexascale compute project , you can code and test it on a Mac pro importing/translating whatever code you need from C/C++/Objective-C/Python and maybe later C# too, and later with simple script migrate it to production clusters running Linux or BSD or whatever where you have the open-source Swift 3 compiler.

[Insert Tim Cook voice in keynote mode]
That's Awwwnmaaaaaahhhzzinnnnng
[End of Tim Cook voice keynote mode]

Swift is a non issue. It's as relevant as Objective-C was. If you want a true open source alternative to C# try Vala. Beside why would you migrate your python application to swift??? People use python where high performance isn't an absolute necessity else they code their application in C/C++ or Rust. If your software architect is dumb enough to chose an interpreted language for a high performance application from inception then you should replace him by a competent one.
 

Mago

macrumors 68030
Aug 16, 2011
2,789
912
Beyond the Thunderdome
Swift is a non issue. It's as relevant as Objective-C was. If you want a true open source alternative to C# try Vala. Beside why would you migrate your python application to swift??? People use python where high performance isn't an absolute necessity else they code their application in C/C++ or Rust. If your software architect is dumb enough to chose an interpreted language for a high performance application from inception then you should replace him by a competent one.
Python is widely used in hpc clusters, but complied thru pypy or others where it's performs at about 80-90% of the same routine in C, further scipy and numpy always run in compiled code as well pyCuda and pyCL.

May seems counter to logic but most researchers prefer to wait 5hr instead 4 if they can code everything on python, go to hpcwire.com and get informed,seems you lost the python bus just when in converted into AirBus.
 
Last edited:

pat500000

Suspended
Jun 3, 2015
8,523
7,515
Two possibilities, either their hardware is so strong and performant that it will defend itself, or... their Marketing department completely sucks.

One thing is waiting to adjust clocks on their hardware to counter what your competitor has to offer, second thing is being completely silent, and letting all the attention go to your competitor...[/Q
Do you think they would adjust it at this time? Hopefully they will
[doublepost=1462806939][/doublepost]Oops I meant to ask will amd be able to adjust it at this time?
 
Status
Not open for further replies.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.